The present invention relates generally to in-memory data grids, and more specifically, to identifying reference data, creating a reference map for the reference data, and replicating the reference map across multiple partitions of an in-memory data grid.
A contemporary in-memory data grid (IMDG) partitions and stores data in memory across an elastic pool of interconnected servers. Data partitioning provides parallelism and enables an IMDG to scale its capacity. For example, in response to an increase in data, the IMDG may simply add servers to handle the growing workload. The partitioned data is then typically load-balance by re-distributing the partitioned data across the newly added servers on which the IMDG is hosted. In a contemporary IMDG, the data may be stored redundantly on multiple servers to ensure high availability of the data in case a server or network link fails.